DC to Microwave Characterization and Modeling of the Cryogenic Performance of Low-Noise HEMT's

We extensively investigated the effects of the temperature on the DC behaviour, the small signal and the noise performance at microwave frequencies of pseudomorphic HEMTs. The measured data were then employed to extract temperature dependent models by means of several extraction techniques. The relevant experimental data show that the most important electrical parameters, such as the output current, the threshold voltage, the transconductance, the forward transmission coefficient and the noise figure, are sensibly affected by thermal phenomena. The knowledge of the small signal equivalent circuit of microwave GaAs FET’s is a crucial point for the design of low noise amplifiers and is a very useful tool to support the analysis of the transistor performance. In the present work, we report the results of our experimental activity concerning the characterization and the application of several improved procedures for the extraction of the model element values from DC, scattering (S-) parameter and noise figure measurements. The developed modeling procedures are: direct extraction, neural network techniques and evolutionary algorithm approach. The very good agreement between the simulated and measured parameters confirms the validity of the proposed methods. To carry out the experimental activity, we employed a properly designed cryogenic set-up operating in our laboratory that allows us to perform DC and microwave characterization down to 30 K.